Research map / Video / Video generation
VividCam: Learning Unconventional Camera Motions from Virtual Synthetic Videos
Qiucheng Wu, Handong Zhao, Zhixin Shu, Jing Shi, Yang Zhang, Shiyu Chang
Read the paperCheck your idea against 137,217 papers
AI-extracted. The measured effects below were read by AI from the paper's full text (no summary yet). Numbers and claims can be wrong or missing context, so check the paper before you cite or build on it.
Measured effects reported in the paper
| What was changed | Where | Measured | Effect |
|---|---|---|---|
| add appearance LoRA during training (ablation) | ViVidCam · Human study ablation | Action Correctness: 0.75 -> 0.93 | Improved |
| add appearance LoRA during training (ablation) | ViVidCam · Human study ablation | Realism: 0.60 -> 0.74 | Improved |
| add temporal consistency loss (ablation) | ViVidCam · Human study ablation | Action Correctness: 0.57 -> 0.93 | Improved |
| add temporal consistency loss (ablation) | ViVidCam · Human study ablation | Realism: 0.76 -> 0.74 | Hurt |
| add style-aligned prompts (ablation) | ViVidCam · Human study ablation | Realism: 0.65 -> 0.74 | Improved |
| add style-aligned prompts (ablation) | ViVidCam · Human study ablation | Action Correctness: 0.93 -> 0.93 | No clear effect |
| add A-LoRA, A-LoRA dropped at inference, over only dual adaptation (ablation) | ViVidCam · Human study ablation | Realism: 0.61 -> 0.74 | Improved |
| keep appearance LoRA at inference (ablation) | ViVidCam-Cog · Inference ablation | FVD: 1866.40 -> 2097.61 | Hurt |
| keep appearance LoRA at inference (ablation) | ViVidCam-AC3D · Inference ablation | FVD: 1721.45 -> 2156.02 | Hurt |
| raise temporal loss weight from 0.03 to 0.3 (ablation) | ViVidCam-Cog · Lambda sweep | TransErr: 0.5859 -> 0.4011 | Improved |
| raise temporal loss weight from 0.3 to 3 (ablation) | ViVidCam-Cog · Lambda sweep | FVD: 1866.40 -> 2124.11 | Hurt |
| add RealEstate10K real videos to synthetic training (ablation) | ViVidCam trajectory-based · RealEstate10K-based setting | Action Correctness: 0.91 -> 0.88 | No clear effect |
Closest papers by meaning
- ViVid-1-to-3: Novel View Synthesis with Video Diffusion Models (CVPR 2024 · similarity 0.90)
- AC3D: Analyzing and Improving 3D Camera Control in Video Diffusion Transformers (CVPR 2025 · similarity 0.89)
- Reconstruct, Inpaint, Test-Time Finetune: Dynamic Novel-view Synthesis from Monocular Videos (NeurIPS 2025 · similarity 0.88)
- VD3D: Taming Large Video Diffusion Transformers for 3D Camera Control (ICLR 2025 · similarity 0.88)
- Training-free Camera Control for Video Generation (ICLR 2025 · similarity 0.88)
- VMC: Video Motion Customization using Temporal Attention Adaption for Text-to-Video Diffusion Models (CVPR 2024 · similarity 0.87)
Writing a paper in this area? scime draft is a free LaTeX editor with AI included, real-time co-authors and a check of your references and venue rules before you submit. It also imports OpenAI Prism projects.
